Conduct adversarial code reviews on JAX package changes with multiple models.
npx skills add https://github.com/xshi19/normix --skill interrogate-xshi19
Please help me install this Agent Skill. Skill: interrogate Source: https://github.com/xshi19/normix/tree/main/.cursor/skills/interrogate Command: npx skills add https://github.com/xshi19/normix --skill interrogate-xshi19
This Skill provides a multi-model adversarial review of code changes, ensuring high confidence in the quality and correctness of changes made to critical sections of the normix package.
Run the 'interrogate' skill on the latest code changes to ensure they meet quality standards.
High-intent search queries and answers about installing and using this skill.
Adversarial code review for JAX changes uses multiple models to critique critical sections and complex logic. It defines scope, clarifies intent, spawns diverse reviewers, synthesizes findings, and leads final judgment to ensure correctness.
Multi-model adversarial review combines diverse model perspectives to catch potential issues in critical code. You need it when making complex logic changes to critical package sections, ensuring high confidence in quality and correctness.
Custom review panels for code changes offer default and strong configurations tailored to complexity. You can select the appropriate panel to match the criticality of your package modifications and ensure thorough feedback.
Adversarial code review steps include scope definition, intent clarification, spawning multiple reviewers, synthesizing structured findings, and leading a final judgment. This structured process ensures comprehensive quality feedback on changes.
Multi-model code review works without external dependencies. The skill relies entirely on its internal scripts and references to conduct structured reviews and provide feedback on JAX package modifications.
Adversarial review for normix updates leverages multiple models to challenge code changes, unlike standard reviews. This approach catches hidden issues in critical sections by synthesizing diverse perspectives into a final judgment.
tense-iOrchestrate adversarial code reviews with opposing-model reviewers and synthesized verdicts.
KangJiSseokParse review-comments.md to apply accepted code changes and run QA.
trevorbyrumOrchestrate multi-model code reviews with static analysis into a unified report.
makinuxRun multi-model adversarial reviews with cross-critique and calibrated synthesis.
